Exception: User API key and team id and user id missing from custom callback.

[ovizii]

I am just taking my first steps with litellm so please bear with my noob questions.

I am trying to debug a problem, but my logs are just overwhelming me, so I started getting rid of the issues from the logs one by one.

I am running litellm via a docker-compose file and have created my own config.yaml
I only added 3 OpenAI models, no routing, nothing, trying to stay pretty basic.

Currently, I see this error repeating almost every second in my logs. Google did not yield any results for me. :-(
I mean I do understand what the error is saying, but obviously my API key is set and requests are actually working. No idea where or why I would need to set a team and user ID?

litellm  | Exception: User API key and team id and user id missing from custom callback.
litellm  | 18:46:40 - LiteLLM Proxy:ERROR: proxy_track_cost_callback.py:135 - Error in tracking cost callback - User API key and team id and user id missing from custom callback.
litellm  | Traceback (most recent call last):
litellm  |   File "/usr/lib/python3.13/site-packages/litellm/proxy/hooks/proxy_track_cost_callback.py", line 104, in _PROXY_track_cost_callback
litellm  |     raise Exception(
litellm  |         "User API key and team id and user id missing from custom callback."
litellm  |     )

[yigitkonur]

same here! langfuse and sentry configured as callback.

[ishaan-jaff]

are you calling health check @ovizii / @yigitkonur ?

[ovizii]

I didn't.

[yigitkonur]

yes, I do!

[lazariv]

@ishaan-jaff Yes, same error here when calling health check. SpendLogs in DB disabled, and S3 configured as callback

[lazariv]

#9585

[Yuhui0620]

Similar problem here when litellm.acompletion was called inside custom_llm_handler.acompletion (follow the doc here).

code:

class DeepSeekCascadeLLM(CustomLLM):
    # outer acompletion call
    async def acompletion(
            self,
            api_base: str,
            api_key: str,
            model: str,
            messages: list,
            optional_params: dict,
            litellm_params=None,
            **kwargs
        ) -> litellm.ModelResponse:
        ...
        # internal acompletion call
        deepseek_response = await litellm.acompletion(
            model=f"deepseek/{model}",
            messages=messages,
            base_url=api_base,
            api_key=api_key,
            extra_headers=auth_headers,
            **passthrough_body
        )

An error was raised after the internal call finished:

06:39:45 - LiteLLM Proxy:ERROR: proxy_track_cost_callback.py:136 - Error in tracking cost callback - 'NoneType' object has no attribute 'get'
Traceback (most recent call last):
  File "/usr/lib/python3.13/site-packages/litellm/proxy/hooks/proxy_track_cost_callback.py", line 40, in _PROXY_track_cost_callback
    user_id = cast(Optional[str], metadata.get("user_api_key_user_id", None))
                                  ^^^^^^^^^^^^
AttributeError: 'NoneType' object has no attribute 'get'

I tried to figure it out and found that the kwargs of _PROXY_track_cost_callback is not the same in the outer and internal acompletion invocation, the former is missing some parameters (e.g. metadata) which causes the error above.

• internal litellm.acompletion debug message:

06:39:45 - LiteLLM Proxy:DEBUG: proxy_track_cost_callback.py:31 - completion kwargs: {'litellm_trace_id': None, 'litellm_call_id': 'c1eafa6f-e75b-4b1b-a611-b3a2cf58d662', 'input': [{'r
ole': 'user', 'content': 'hello'}], 'litellm_params': {'acompletion': True, 'api_key': 'sk-anything', 'force_timeout': 600, 'logger_fn': None, 'verbose': False, 'custom_llm_provider': 
'deepseek', 'api_base': 'http://192.168.4.128:9050/ds-gateway-test/v1/chat/completions', 'litellm_call_id': 'c1eafa6f-e75b-4b1b-a611-b3a2cf58d662', 'model_alias_map': {}, 'completion_c
all_id': None, 'aembedding': None, 'metadata': None, 
...

• outer custom_llm_handler.acompletion debug message:

06:39:45 - LiteLLM Proxy:DEBUG: proxy_track_cost_callback.py:31 - completion kwargs: {'litellm_trace_id': None, 'litellm_call_id': 'cb8a1ff1-1b4b-4a96-b800-41d7dfc9d952', 'input': [{'r
ole': 'user', 'content': 'hello'}], 'litellm_params': {'acompletion': True, 'api_key': 'sk-anything', 'force_timeout': 600, 'logger_fn': None, 'verbose': False, 'custom_llm_provider': 
'deepseek-cascade-llm', 'api_base': 'http://192.168.4.128:9050/ds-gateway-test/v1', 'litellm_call_id': 'cb8a1ff1-1b4b-4a96-b800-41d7dfc9d952', 'model_alias_map': {}, 'completion_call_i
d': None, 'aembedding': None, 'metadata': {'requester_metadata': {}, 'user_api_key_hash': '88dc28d0f030c55ed4ab77ed8faf098196cb1c05df778539800c9f1243fe6b4b', 'user_api_key_alias': None
, 'user_api_key_team_id': None, 'user_api_key_user_id': 'default_user_id', 'user_api_key_org_id': None, 'user_api_key_team_alias': None, 'user_api_key_end_user_id': None, 'user_api_key
': '88dc28d0f030c55ed4ab77ed8faf098196cb1c05df778539800c9f1243fe6b4b', 'user_api_end_user_max_budget': None, 'litellm_api_version': '1.61.0', 'global_max_parallel_requests': None, 'use
r_api_key_team_max_budget': None, 'user_api_key_team_spend': None, 'user_api_key_spend': 0.0, 'user_api_key_max_budget': None, 'user_api_key_model_max_budget': {}, 'user_api_key_metada
ta': {},
...

Maybe some extra parameters need to be passed in the internal litellm.acompletion? But I can not find the proper way to do.

[Yuhui0620]

oh, i missed the metadata arg (here) of litellm.acompletion, it worked normally after litellm_params["metadata"] was passed to litellm.acompletion.

[alois-nejo]

may i ask where you fixed this? can this be set in the config.yaml?

[Yuhui0620]

pass litellm_params['metadata'] to the internal accompletion method, like this:

 async def acompletion(
        self,
        model: str,
        messages: list,
        api_base: str,
        api_key,
        optional_params: dict,
        litellm_params=None,
        **kwargs
    ) -> ModelResponse:
        return await litellm.acompletion(
            ...,
            metadata=litellm_params['metadata']
        )

[dotmobo]

Hi, i have the same problem when i call the healthcheck from the admin page (docker image v1.72.6-stable):

{"message": "Error in tracking cost callback - User API key and team id and user id missing from custom callback.", "level": "ERROR", "timestamp": "2025-06-25T09:07:30.998802", "stacktrace": "Traceback (most recent call last):\n File "/usr/lib/python3.13/site-packages/litellm/proxy/hooks/proxy_track_cost_callback.py", line 178, in _PROXY_track_cost_callback\n raise Exception(\n "User API key and team id and user id missing from custom callback."\n )\nException: User API key and team id and user id missing from custom callback."}

How do you fix it ? thanks